Part 1:

Personally, I don’t think that AI can ever replace economists, because humans acquire the know-how behind creation of economic models, policy making and tangible interventions. AI is still very useful and is already transforming the field economics. It is used in a lot of practices such as, data analysis, coding, creation of economic models and mathematical formula. The only thing that AI cannot fully grasp is policy making and behavioural interventions.

There is field in economics called behavioural economics. It’s a mixture of psychology and traditional economics. AI has become very advanced in the calculations belonging to traditional economics, but it has not yet grasped the behavioural concepts and is unable to comprehend them. Of course, it understands the concepts in theory but cannot interpret them properly into policy making. In my opinion, I don’t think it will be able to do that anytime soon, because human interactions, personal evaluation and critical thinking cannot be simply put into a database.

I think that AI will be used more commonly in the government sector to help with taxation and data analysis, but policy making will remain to be a human responsibility. In the long run, AI may become more emotionally intelligent, which will make it more useful in policymaking. It is already being used in many statistical aspects of economic analysis to form matrixes, assumptions and models. Moreover, it is starting to be used in academic contexts to make coding short cuts, which makes reaching conclusions faster and more efficient. One of the main goals in economics is increasing efficiency, and I believe that AI is already establishing that. It also improved welfare by improving resource allocation and minimizing wastage of resources. AI has already proven successful in coding of mathematical formula for graphical creation and data forecasting, which is critical in creation of economic models, so who knows what else it might master in the future.

Part 2:

(copilot) AI chat link:https://copilot.microsoft.com/shares/VpZoPwqEEZepkQWVGtvco

Prompt: Suggest ways in which AI will be used in the field of economics. It can be in a professional/working world or in an academic context.

Part 3:

These are the interview questions for part 3 and 4:

1.⁠ ⁠How do you see artificial intelligence currently influencing economic decision-making, both at the individual and policy level?

2.⁠ ⁠In your opinion, what distinguishes genuine economic value from what some might call “AI hype”?

3.⁠ ⁠Can you think of examples where AI is being overestimated in economic contexts?

4.⁠ ⁠Why do you think AI is often presented as a “solution” to complex economic problems?

5.⁠ ⁠What risks come from relying too heavily on AI-driven economic models or predictions?

6.⁠ ⁠How should economists critically evaluate claims about AI’s effectiveness?

7.⁠ ⁠When AI is used in economic policy (like welfare systems or pricing), who should be held accountable for the outcomes?

8.⁠ ⁠⁠    How do issues like bias or lack of fairness affect the credibility of AI in economics?

9.⁠ ⁠⁠    Looking ahead, how do you think AI will shape the future of labour markets and economic inequality, especially for younger generations?

I conducted the interview with my Microeconomics professor, Dr. Ahmed Tohamy. I took personal notes during the interview, but we recorded the interview using Zoom and used its AI transcription feature. So, I will also include it along with my notes. The AI summary is quite long, but it included some critical points. If you are interested in the topic, you can read it, but my notes pretty much summarize everything.

Recorded interview link: https://aucegypt.zoom.us/rec/share/7OPKq7kVArK16zlnJfrO0JbUElIQA3H_-vZnxfCK343Z48w9RFP80eoM4-lvaFCH.65YCB9SXFYsh3q17

My personal notes:

(They were very messy and were written in incomplete sentences. I cleaned them up and put them in chronological order. I used AI here to correct some grammatical mistakes and compress my notes)

1) AI is already affecting everyday decisions. Students use it just to get assignments done, people use it for cover letters, and banks use AI assistants. It’s also used in logistics and customer service, making things more streamlined, faster, and cheaper at scale. On a bigger level, governments use AI to predict things like inflation and exchange rates. Some countries are even starting to give AI more decision-making power, which is concerning. The main issue is that people are starting to rely on AI without understanding how it actually works, so decision-making is being externalised. 

2) Real value from AI comes when it actually creates new services, opens markets, or reduces costs. But a lot of it can also be hype. There are concerns about data access, who owns it, and how it is used. For example, selling data about populations or using it in ways that benefit other countries more than your own. So, it’s not always clear if AI is truly adding value or just being presented that way.

3) People overestimate AI because they think it can fully replace thinking. But in reality, we don’t know how it makes decisions since it’s like a black box. There are extreme examples, like AI simulations making unrealistic or extreme decisions in war scenarios. This shows that AI doesn’t always think in a way that aligns with human judgment, especially in complex situations.

4) AI is often presented as a solution because it is fast, efficient, and works at scale. It makes things look easier than they actually are. But the issue is that it can oversimplify problems. Instead of actually solving them, it sometimes just replaces the thinking process. People rely on it instead of engaging deeply, which again links to the idea of externalising decision-making. 

5) One of the biggest risks is that AI is a black box, so we don’t understand how it reaches conclusions. There is also the issue of “garbage in, garbage out,” meaning bad data leads to bad decisions. This is especially risky in things like welfare systems or tax auditing, where decisions directly affect people’s lives. There’s also a risk of blindly trusting AI outputs without questioning them.

6) Students and economists shouldn’t just use AI to tick a box or pass a course. Instead, they should focus on how they use it and how they ask questions. AI should be understood and audited, not just used blindly. It can help with repetitive or simple tasks, but human thinking is still needed for deeper understanding and analysis.

7) When AI is used in policies like taxation or welfare, it becomes unclear who is responsible for the outcomes. For example, if an AI system makes a wrong decision, is it the government, the developer, or the system itself? This lack of clear accountability is a big issue, especially when decisions impact people directly.

8) AI depends heavily on data, so if the data is biased, the outcomes will be biased too. This affects how fair and credible AI systems are, especially in areas like welfare targeting. Again, it comes back to the idea that bad input leads to bad results.

9) AI is expected to take over around 30 to 40 percent of current tasks. This puts many jobs, especially white-collar jobs, at risk. This situation is similar to the Industrial Revolution, where automation replaced workers. The problem back then was that the benefits were not distributed fairly, and the same thing could happen again. If there is no policy intervention, more people could become poorer while those who own AI technologies gain more. There is also strong global competition to build better AI, which makes regulation harder. At the same time, there is a huge push in infrastructure and energy to support AI growth. 

Zoom’s AI transcription feature:

This interview was conducted as part of an AI literacy course and intercultural learning project, where Rokaya interviewed Dr. Ahmed about artificial intelligence’s influence on economic decision making. Dr. Ahmed discussed how AI is currently used at individual and policy levels, including applications in learning, job searching, banking, and government predictions, while highlighting both benefits and risks such as data externalization and decision-making delegation. They explored concerns about AI replacing economists and other professionals, with Dr. Ahmed sharing his experience using Claude AI to assist with economic modelling research, which produced results identical to his own work but at a much faster pace. The conversation concluded with a discussion about historical technological disruptions and their impact on labour markets, comparing current AI developments to the Industrial Revolution and considering potential policy interventions to address economic displacement.

AI’s Impact on Economic Decisions

Ahmed and Rokaya discussed the impact of artificial intelligence on economic decision-making at both individual and policy levels. Ahmed explained how AI is used for learning, job searching, and financial planning, while highlighting both its benefits and risks, such as the externalization of decision-making power. They also touched on AI’s use in government for predicting economic factors and its controversial application in military contexts, including the potential for AI to advise on the use of tactical nuclear weapons, which raised concerns about the implications of delegating critical decisions to AI systems.

AI Applications and Implementation Challenges

Ahmed discussed the applications and limitations of AI models in various contexts, including military strategy and civilian applications like tax compliance and welfare targeting. He emphasized the importance of evaluating AI value based on creating new services, reducing costs, and managing risks, while highlighting potential issues with data quality and privacy. Ahmed also touched on the risks associated with data access and national sovereignty when implementing AI systems.

AI Data Cross-Border Challenges

Ahmed and Rokaya discussed the challenges and risks associated with sharing AI data across countries, particularly focusing on data security and the potential for exploitation. Ahmed emphasized the need to balance creating value through new services and products with reducing operational costs while managing risks. They also touched on the importance of ensuring that AI models developed in different countries are tailored to the specific needs and interests of their populations rather than being influenced by global or Western interests.

AI-Driven Economic Models Discussion

Rokaya and Ahmed discussed the risks of relying heavily on AI-driven economic models, particularly the issue of externalizing decision-making processes without full understanding of the underlying decisions. Ahmed shared an example of how AI tools could create problems in personal situations, while Rokaya expressed her preference for learning through hands-on practice rather than using AI to complete assignments. The conversation touched on how students approach their studies, with Rokaya emphasizing the value of engaging with course material independently, while Ahmed noted that some students may treat courses as mere “tick box exercises” to meet minimum requirements.

AI in Academic Assignments Discussion

Ahmed and Rokaya discussed the use of AI in academic assignments, particularly in economics courses. Ahmed shared his perspective on AI’s impact on education and the labour market, noting that while AI can perform mundane tasks, students need to understand its limitations and potential effects on job roles. He emphasized the importance of teaching students to audit and understand AI outputs rather than strictly prohibiting their use. The conversation touched on the broader implications of AI technology for the future of work, though it ended abruptly as Rokaya began to ask about AI replacing economists.

2008 Housing Market Crisis Analysis

Ahmed and Rokaya discussed the 2008 housing market crisis, which affected global financial sectors including Dubai, the UK, and Europe. Ahmed explained that while the crisis started with a UK bank collapse, constructing an economic model to analyse the causal mechanisms is a complex task that requires comparing different scenarios. Ahmed mentioned working with a colleague over the winter break to develop a specific economic model for this context, though he noted that creating such targeted models is not common among economists.

AI Code Writing Experience

Ahmed discussed his experience with an AI model that can write and edit code autonomously, comparing its efficiency to his own work. He noted that while the model produced identical results to his previous work in less time, the high cost of the subscription ($200 per month) makes it impractical for his current income level. Ahmed also mentioned the emerging concept of “vibe coding,” where AI generates code based on given instructions, and highlighted the growing importance of programming and computer science in the future workforce.

AI Growth and Data Challenges

Ahmed discussed the rapid growth in code submissions and data storage challenges driven by AI tools like ChatGPT, noting that while AI can speed up implementation and analysis, human expertise remains valuable for strategic thinking and asking insightful questions. He compared current technological changes to historical innovations like the Industrial Revolution, explaining how past bursts of innovation led to significant growth and global impact. Rokaya confirmed that Ahmed’s responses addressed all her questions and expressed satisfaction with the discussion.

AI Impact on Labor Markets

Ahmed discussed the historical impact of automation on labour markets, explaining how factory workers in the 1980s lost their jobs to machines and how this led to the rise of right-wing politics. He warned that similar disruptions are happening now with AI technology, which could affect 30-40% of current jobs, particularly white-collar positions held by middle to upper classes. Ahmed expressed uncertainty about whether governments will intervene to redistribute AI-generated wealth, noting the international competition that could prevent such policies. The conversation concluded with a discussion about using AI to grade assignments, though Ahmed noted practical challenges with implementing this at scale.

Part 4:

Questions:

1.⁠ ⁠How do you see artificial intelligence currently influencing economic decision-making, both at the individual and policy level?

2.⁠ ⁠In your opinion, what distinguishes genuine economic value from what some might call “AI hype”?

3/4/5. (the three questions were contracted together)

Why do you think AI is often presented as a “solution” to complex economic problems? Distinguish and explain its risks.

6.⁠ ⁠How should economists critically evaluate claims about AI’s effectiveness?

7.⁠ ⁠When AI is used in economic policy (like welfare systems or pricing), who should be held accountable for the outcomes?

8.⁠ ⁠⁠    How do issues like bias or lack of fairness affect the credibility of AI in economics?

9.⁠ ⁠⁠    Looking ahead, how do you think AI will shape the future of labour markets and economic inequality, especially for younger generations?

Alia Abdelhamid is my relative. She now owns her own consulting business.

She started her career doing research for the bank of England, then progressed into consulting.

Her degree was BSc and MSc in Economics and Economic Policy.

Interview Link: https://drive.google.com/file/d/1-jtHNAxwVRi5LR4EOMNIB5NTKw1tSuLI/view?usp=drivesdk

My Personal Summary:

(They were very messy and were written in incomplete sentences. I cleaned them up and put them in chronological order. I used AI here to correct some grammatical mistakes and compress my notes)

AI is very impactful in economic decision-making, especially at a big-picture level. Since economics is about interpreting data, AI helps analyse much larger datasets faster and with more advanced methods. It improves how decisions are made but doesn’t replace them. AI can create real value, but it’s also overhyped. It’s not as smart as people think and depends a lot on how it’s used. Bad prompts or weak data can completely distort results, so validation and data quality really matter.

People overestimate AI because it looks like it can solve complex economic problems quickly. In reality, economics is too complex for models alone. AI simplifies things, which makes it seem objective, but that can create false confidence, especially if assumptions or data are weak.

To evaluate AI properly, economists need to be clear about their data, methods, and assumptions, and explain why they’re using it in the first place. It’s not just about using AI, but it’s about understanding it.

When AI is used in policy, humans should still be accountable for policy decisions. Researchers handle the models and data, while policymakers are responsible for the final decisions. AI is just a tool, and without proper checks, it can create false confidence. 

Bias is a big issue because AI depends on data. If the data is biased, the results will be too, which can reinforce inequality instead of fixing it and affects how credible AI is. 

In the future, AI will likely automate routine tasks, increasing demand for high-skill jobs and putting pressure on lower-skill ones. For younger people, this means more tools but also more competition. It could also help identify labour market gaps and push universities to adapt.

AI is a tool that can improve efficiency and innovation, but it can’t replace human thinking. Its value depends on how well people actually understand and use it. 

These are her answers to my questions:

The interview was recorded, but she also wrote down the answer to each question. Fortunately, she enjoyed the process and told me that I came up with some smart questions!

(Questions are written in first person because she answered them herself, but she still did an interview with me during her free time to clarify the points further)

1.⁠ ⁠AI & economic decision-making

I see AI on an aggregate level as highly impactful for decision makers. At its core, economic decision-making is about interpreting data and how that data is utilized and interpreted. AI allows policymakers and researchers to use more advanced econometric methodologies, refine data further, analyse much larger datasets more efficiently, and move beyond traditional random sampling. This improves the depth and scale of analysis. Overall, it enhances how economic decisions are informed, rather than replacing them.

2.⁠ ⁠Value vs AI hype

AI is a powerful but also dangerous tool if it is not used properly. It is not as “smart” as it may seem and needs correct utilization to reflect robust results. For example, wrong prompting can directly affect modelling and therefore the outputs. Additionally, if the data sources are not trusted or verified, AI may unintentionally rely on faulty or low-quality information. That’s why validation and data integrity are critical.

3.4.5.

AI is often overestimated in economic contexts where it is seen as something that can quickly unlock or solve global economic crises. In reality, economics is a complex science backed by constantly evolving research, ongoing data updates, and exposure to global shocks, which cannot be fully captured by models alone.

This is also why AI is often presented as a “solution” to complex economic problems—it simplifies complexity into clear outputs and is often seen as objective or neutral, which makes it very appealing for policymakers and decision-makers.

The risk of relying too heavily on AI-driven economic models is that it can create false confidence in results that may be based on weak assumptions, poor data quality, or lack of context. Ultimately, it can mislead decision-making if not critically assessed and interpreted by economists with the right expertise.

6.⁠ ⁠How should economists critically evaluate claims about AI’s effectiveness?

Critically assessing claims about AI’s effectiveness is quite simple in principle. Economists must provide clear proof behind all data used to ensure robustness of the results. There also needs to be well-explained reasoning for using specific AI tools, and how they fit within the methodology being applied. This should be supported by relevant econometric techniques and the existing literature behind the hypotheses. Ultimately, it comes down to transparency in data, method, and assumptions to properly validate any claim.

7.⁠ ⁠Accountability in AI-driven policy

When using AI in policymaking, such as welfare systems or pricing, the accountability should sit with both the research teams and the policymakers. AI is a tool, and as I mentioned earlier, it is only as strong as the data, methodology, and how it is applied. Research teams are responsible for ensuring robustness in the models, data quality, and the correct use of methodologies. Policymakers are then responsible for how these outputs are interpreted and implemented in real-world decisions. Ultimately, because AI can create false confidence if not properly validated, accountability must remain fully human at both stages.

8.⁠ ⁠Bias, fairness, and AI credibility in economics

It all circles back to data and how models are constructed. Good economists build models for policymaking that are robust against bias, and that starts with using relevant, representative, and reliable data.

When AI systems are trained on biased or incomplete data, they can unintentionally reinforce existing inequalities rather than correct them. This directly affects credibility, because the outputs may look precise but are not necessarily fair or accurate in an economic sense.

Ultimately, fairness and bias are not just ethical concerns—they affect the quality and reliability of economic decision-making itself.

9.⁠ ⁠AI, labour markets, and inequality

Looking ahead, I think AI will significantly reshape labour markets, especially through the automation of routine and repetitive tasks. This will likely increase skill polarization, where demand rises for highly skilled, analytical roles, while pressure builds on lower-skilled or entry-level jobs.

For younger generations, this creates both opportunity and risk. There is greater access to tools that can enhance productivity, but also higher competition and a more demanding skills threshold to enter the job market.

At the same time, AI might also help identify the gaps between labour market demand and supply more clearly. In the long run, this can guide universities into restructuring certain courses and adapting curricula to better fit evolving economic trends.

Conclusion

I would like to conclude by stressing that ultimately, it is important to understand that AI is there to be utilised as a tool to increase efficiency, creativity, and innovation. It is not a human replacement. Without the knowledge of how to correctly use and interpret AI, its value is significantly limited, especially in economic decision-making and policy contexts.

Part 5:

Personally, I wasn’t surprised with my findings. They are similar to what I initially thought of and to what my interviewees said. AI can’t replace economists, and this is what we all think. However, what stood out to me more was how both interviews really questioned the hype around AI.Before this, I knew AI was useful, but I didn’t fully think about how much people actually trust it without understanding how it works. My microeconomics professor proposed that the AI is like a “black box” and that it’s based on the data it’s fed. He used the phrase “garbage in, garbage out” on AI and I think that this is the most relatable comment I have heard. AI is very biased and is based on western data, when it comes to economics. It needs to be broader and consider the economy of Arab, African and Asian countries when making, calculations, models and assumptions.  We don’t fully know how AI reaches its decisions and that is a bit concerning. Bias can definitely create inequality, which is main cause of market failure.

What I found exciting is how powerful AI is when it is used properly. It can analyse large amounts of data, make processes faster and improve efficiency, which is very important in economics.  However, I’m worried about how people sometimes blindly rely on it and think that is trustworthy. A lot of people are starting to depend on AI instead of thinking for themselves, especially in academic work. This is kind of linked to the idea of externalising decision making. In the interviews, we reached a conclusion that AI is useful, but can’t be used to externalize decision making, especially in economics. In policy making, human knowledge and thinking is always required, as AI can’t grasp some concepts and won’t match human thinking skills. We created AI, so we are definitely smarter and have more to offer than we think we are capable of.

Another important point is that AI can create false confidence. Just because something is data driven or looks technical does not mean it is correct. Economics is complex and AI can sometimes oversimplify problems and make them seem easier than they are. I think my interview questions were strong because they focused on both the benefits and risks of AI, not just one side. They also challenged the idea that AI is always a solution, which helped bring out more realistic answers instead of exaggerated claims. The questions were open ended and not leading, which was good because they allowed the interviewees to explain their thinking, especially on topics like AI hype, bias and accountability. I learned a lot of new information that will definitely be useful in my future career prospects. I really did enjoy the conversations, especially with my professor. This assignment gave me great opportunity to ask questions to deepen my knowledge even further on my field of study. I also added a lot of follow up question because I was actually interested in the information, they were providing me with, not just for the sake of the assignment.

Right now, I think I need to be more careful using AI and appreciate what it has to give. AI will definitely help make economists in various ways such as data analysis etc. However, humans will always be needed to monitor and control AI outcomes. It is definitely a useful tool in the field, but it needs to be used wisely and with precision, as bias is very common and this can result catastrophic market failures all over the world. Instead of just using it to make things easier, I should focus on understanding it and questioning the results it gives me. This made me realise that AI is a very useful tool, but it should not replace human thinking.